JP5575270B2 - Ad source and keyword set adaptation in online commerce platforms - Google Patents

Description

CROSS-REFERENCE Other applications are incorporated herein by reference for all purposes, the entitled January 13 dated filed 2010 "RETRIEVAL METHOD AND SYSTEM THEREOF FOR ONLINE TRADING PLATFORM ( online commerce Claiming priority based on Chinese Patent Application No. 201010003928.7 ”.

  The present invention relates to the field of website data processing, and more particularly to management of Internet advertisements.

  An online commerce platform is a computer information system provided by a platform service provider to facilitate online commerce. An online commerce platform can implement online commerce between sellers and buyers. On an online commerce platform, potential buyers can search for (retrieve) advertising sources provided by sellers by entering specific keywords. Under the bidding ranking model on the Internet, the association between the advertising source provided by the seller (eg text file, image, etc.) and the corresponding keyword (ie purchased word or unit) Is called a match (match) between a keyword and a keyword. A seller can display an advertisement that includes one or more advertisement sources by paying for a particular keyword. For example, if the seller has paid for the keyword “book” and the search word entered by the potential buyer is associated with the keyword “book” under certain rules, the advertising source specified by the seller (eg, , Text files, images, etc.) will be displayed to potential sellers. Sellers benefit from such a match between keywords and ad sources. Activities that benefit the seller, such as click-through advertisements by potential buyers, inquiries by potential buyers, or transactions with potential buyers, will be referred to as seller profits.

  In existing online commerce platforms, when a potential buyer enters a set of query keywords, the advertising source provided by the seller is searched based on a particular match. Relevance is typically based on how new the ad source is (the most recent ad source is usually preferred) and whether the fit between the ad source and the keyword can benefit the ad source Generated. For example, a seller has two advertising sources, “Various mobile phone sales” and “Imported mobile phone sales”, which have three keyword sets “mobile phone”, “Nokia mobile phone”, and “ Associated with "Apple Mobile Phone". Assuming that each advertising source can only be associated with at most two keyword sets, different degrees of relevance can yield the following profit levels to the seller on a scale of 1-10. When “Sale of various mobile phones” is adapted to “mobile phone”: 10, When “Sale of various mobile phones” is adapted to “Nokia mobile phone”: 7, “Sale of various mobile phones” is “ When "Apple mobile phone" is adapted: 6, "Sales of imported mobile phone" is "Adapted to mobile phone: 9", When "Sales of imported mobile phone" is adapted to "Nokia mobile phone": 7. “Sales of imported mobile phone” is adapted to “Apple mobile phone”: 5. For matches between advertising sources and keyword sets generated according to existing technology, the advertising source “Various mobile phone sales” is adapted to the keywords “mobile phone” (profit 10) and “Nokia mobile phone” (profit 7), Assuming that the advertising source “sales of imported mobile phones” is matched to the keyword “Apple mobile phones” (profit 5), that fit would give the seller a total profit of 22 (ie 10 + 7 + 5).

  However, in the above example, “sales of various mobile phones” is related to “mobile phone” (profit 10) and “apple mobile phone” (profit 6), and “sales of imported mobile phones” is referred to as “Nokia mobile phone” ( There is also a match between advertising sources and keyword sets associated with profit 7). This fit will bring the total profit 23 to the seller. In practice, when the number of keyword sets to which the ad source provided by the seller fits is limited, only the latest ad source is considered, or the fit is maximized to the current ad source or a group of similar ad sources The possibility of obtaining a match between ad sources and keywords that does not necessarily represent every product area offered by the seller, when setting fits between ad sources and keywords only based on whether or not I understand that there is. As a result, some of the product areas offered by the seller may have no or little advertising sources that match the keyword set. In such a case, a potential buyer cannot obtain a corresponding appropriate advertising source when entering a keyword set to perform a product search. If a potential buyer repeats the search, the search performance of the online commerce platform server is also degraded. Furthermore, existing methods for generating matches between advertising sources and keywords may not provide the greatest benefit to the seller.

  Various embodiments of the invention are disclosed in the following detailed description and the accompanying drawings.

6 is a flowchart illustrating one embodiment of a process for generating a match between an advertisement source and a keyword set.

2 is a flowchart illustrating one embodiment of a process for retrieving an advertisement source from an online commerce platform configured according to the process 100 illustrated in FIG.

6 is a flowchart illustrating one embodiment of a process for generating a match between an advertisement source and a keyword.

1 is a block diagram illustrating one embodiment of a device for generating a match between an ad source and a keyword.

The block diagram which shows one Embodiment of the search system in an online commerce platform.

FIG. 4 is a block diagram illustrating another embodiment of a system for generating a match between an ad source and a keyword.

  The present invention is a process, apparatus, system, composition of matter, computer program product embodied on a computer readable storage medium, and / or processor (stored in and / or stored in a memory connected to a processor). A processor configured to execute the provided instructions) and can be implemented in various forms. In this specification, these implementations or any other form that the invention may take may be referred to as techniques. In general, the order of the steps of disclosed processes may be altered within the scope of the invention. Unless stated otherwise, a component such as a processor or memory that is described as being configured to perform a task is a general component that is temporarily configured to perform a task at a certain time, or It may be implemented as a particular component that is manufactured to perform a task. As used herein, the term “processor” is intended to refer to a processing core configured to process one or more devices, circuits, and / or data such as computer program instructions.

  The following provides a detailed description of one or more embodiments of the invention with reference to the drawings illustrating the principles of the invention. Although the invention has been described in connection with such embodiments, it is not limited to any embodiment. The scope of the invention is limited only by the claims and includes many alternatives, modifications, and equivalents. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. These details are for the purpose of illustration, and the present invention may be practiced according to the claims without some or all of these specific details. For the purpose of simplicity, technical material that is known in the technical fields related to the invention has not been described in detail so that the invention is not unnecessarily obscured.

  The present application relates to various general purpose or special purpose computing device environments or configurations including personal computers, server computers, handheld or portable devices, flat panel devices, multiprocessor devices, etc., and distributed including devices as described above. It may be implemented in a computer environment.

  This application may be described in the general context of computer-executable instructions (eg, program modules) being executed by a computer. Generally, program modules may include routines, programs, objects, assemblies, data structures, etc., for performing particular tasks or implementing particular abstract data types. The present application may be implemented in a distributed computing environment. In a distributed computing environment, tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules may be stored on a local or remote computer storage medium, such as a storage device.

  In the embodiment described herein, the seller provides an advertising source corresponding to a particular category. Multiple advertisement sources are obtained according to the distribution of advertisement sources provided by the seller under each category. Since advertising sources are obtained according to the distribution of advertising sources in these categories, it is preferable that the selected advertising sources cover as much of the product field offered by the seller as possible. Corresponding keyword sets are generated based on these advertising sources. Linear programming techniques are used to look for a substantially optimal match of the selected advertising source and keyword set. A programming model is established to include an objective function and a set of constraints. An objective function solution is obtained using a search-based technique. A substantially optimal match between the ad source and the corresponding keyword to be matched is obtained. The resulting fit may represent the seller's maximum profit under certain constraints. The potential buyer can then search for the most relevant advertising source provided by the seller based on the fit established above when entering a keyword set to search for the desired product. Thus, both the potential buyer buying experience and the seller selling experience are improved. This technology also avoids the problem of repeated searches caused by matching between ad sources and keywords generated by other existing systems, and prevents degradation of online commerce platform server performance.

  FIG. 1 is a flowchart illustrating one embodiment of a process for generating a match between an advertisement source and a keyword set. Process 100 may be used in various general purpose or special purpose computing device environments or configurations including personal computers, server computers, handheld or portable devices, flat panel devices, multiprocessor devices, etc., and distributed including devices as described above. May be implemented and executed in a distributed computer environment.

  In step 101, an advertisement source provided by a particular seller and satisfying a predetermined condition is selected. In some embodiments, this selection is made in relation to the category of the advertisement source.

  The advertising source provided by the seller may have a variety of formats, such as text, images, videos, audio files, and any other suitable format suitable for presentation to the user. In some embodiments, advertising sources provided by sellers are classified according to certain rules. For example, the categories of advertising sources may be organized as a tree where the root node is clothing and the sub-nodes include men's clothing, women's clothing, and children's clothing. The sub-node “male clothing” may further include clothing corresponding to different styles, different seasons, and the like. Approaches for classifying advertising sources are well known. Various techniques may be used to classify the advertisement sources provided by the seller as long as the same type of advertisement sources that meet certain criteria are included in the same category. Thus, the advertisement source is selected in relation to each category.

  The predetermined condition specifies which advertising source is selected. For example, the seller may set the number of advertising sources that are matched with keywords before such a match is made. The predetermined condition may require that the number of selected ad sources be greater than the number of ad sources matched with the keywords, so that all ad sources are matched with at least one keyword set. The In some embodiments, ad source selection criteria include when the ad source was created, whether the ad source has already been used before, and so on.

  At step 102, a corresponding keyword set is generated based on the selected advertisement source provided by the seller. Well known techniques such as headword selection and extended matching may be used to generate keywords. The criteria for generating a keyword set requires that the generated keyword set is related to the advertising source at some level and represents the purchase intention of a potential buyer. For example, if the advertising source provided by the seller is the text “Sale of various Nokia mobile phones”, the generated keyword set will be “Mobile phone”, “Nokia mobile phone”, and “Various mobile phones”. May be included.

  In step 103, a planning model is established. The planning model includes an objective function that obeys one or more predetermined constraints. The planning model represents the fit between the selected ad source provided by the seller and the generated keyword set. Matching is sometimes referred to as the association of the advertisement source provided by the seller and the corresponding keyword set.

  In this embodiment, the predetermined constraints include the number of advertisement sources that are matched with the keyword set, the number of keyword sets that each advertisement source corresponds to, and the like. In other embodiments, other constraints may be preset. The value of the constraint is usually preset by the seller. Since the number of ad sources and the number of keywords that each ad source corresponds to are limited, when establishing a planning model that satisfies the above constraints, a specific ad source is associated with a specific keyword set. In order to express whether or not, an integer variable having a value of 0 or 1 is introduced. In this example, an entry having the value 1 represents “correspondence”, and an entry having the value 0 represents “no correspondence”. As will be described in detail below, an objective function for maximizing the seller's profit is set under predetermined constraints. As used herein, seller's profit represents the profit brought to the seller by a match established between a particular advertising source and a particular keyword set. For example, the profit can be measured using the click-through rate for the seller. Hereinafter, detailed processing for establishing a plan model will be described.

  In step 104, the objective function is sought using a search-based solution to obtain a substantially optimal match (correspondence) between the advertising source and the keyword set.

  After the planning model is established in step 103, a search-based solution may be used to solve the planning model objective function, with the result being a substantially optimal fit between the advertising source and the keyword. is there. Details of a method for solving an objective function using a search-based solution will be described below. The total number of advertisement sources in the match is called the number of advertisement sources matched with the keyword set and is pre-specified by the seller. The constraints specified in advance by the seller include the number of advertisement sources that are matched with the keyword set and the upper limit of the number of keyword sets that are matched to each advertisement source.

  In some embodiments, a match between an ad source and a keyword is sent to the seller as it is generated, and the seller determines whether the match meets its requirements and accepts the match. You are allowed to refuse and / or modify.

  The fit solved in this application can be used repeatedly. Once the match between the advertising source and the keyword set is established, it may be used in a variety of applications. In general, when a potential buyer enters a keyword, the online commerce platform server may search for an advertisement source provided by the seller based on a pre-generated match, and the content of the advertisement source to the potential buyer. To return. The fit established in the present embodiment represents the seller's product information, thus increasing the likelihood that a potential buyer will make a purchase through advertising.

  FIG. 2 is a flowchart illustrating one embodiment of a process for retrieving an advertisement source from an online commerce platform configured according to the process 100 illustrated in FIG. That is, in this example, once an advertisement source and keyword set match for a seller is generated, it is stored on the online commerce platform server side.

  In step 201, a query keyword set is entered by a potential buyer.

  In step 202, a corresponding advertisement source is searched based on a match between the keyword set and the advertisement source that was previously established according to the process 100 shown in FIG.

  When a potential buyer enters a keyword set (ie, a desired product or item keyword set), the corresponding advertisement source is searched by the server based on the fit obtained as a result of the process 100. The advertisement source retrieved based on the matches stored in this embodiment accurately reflects the actual product information of the seller and the purchase intention of the potential buyer. Thus, this technique allows potential buyers to find the desired product more accurately without having to repeat the search or change the query keyword. Furthermore, since the advertising sources within the fit established in the embodiments of the present application are selected according to the category, the online commerce platform server need only perform a search according to the category to which the advertising source belongs. Thus, the number of searches performed by the online commerce platform server is reduced, thereby improving the search performance of the online commerce platform server.

  In step 203, the retrieved advertisement source is displayed to potential buyers. In some embodiments, the server configures advertisements to be displayed to potential buyers based on the advertisement source and displays the advertisements on the web page so that the potential buyers By clicking on an advertisement, information about a specific product can be obtained and / or purchased.

  FIG. 3 is a flowchart illustrating one embodiment of a process for generating a match between an advertisement source and a keyword.

  In step 301, advertising sources provided by sellers in current online commerce platform systems are classified according to predetermined classification criteria. Any suitable classification technique may be used. The classified advertisement sources are stored so that an advertisement source can be easily selected from a particular category of advertisement sources later.

  At step 302, a distribution of advertising sources provided by sellers under each category is obtained. In step 303, a predetermined number of advertisement sources are selected from each category according to the distribution.

  The distribution is later used to select advertisement sources in proportion to the number of advertisement sources in each category. By selecting an ad source according to the distribution of ad sources under each category, at least some ad sources from each category are selected and matched to the query without excluding one category. In this way, the appropriate advertisement source provided by the seller is presented to potential buyers, thus increasing the likelihood of user click-through and / or purchase. In some online commerce platforms, positive user feedback enhances the association. Thus, if a potential buyer later enters the same search keyword, the online commerce platform server is more likely to find the same match of the keyword to the advertising source.

  For example, if the distribution of advertising sources provided by sellers under category A and category B is 1: 2, the advertising sources selected from category A and category B also have a ratio of 1: 2.

  In some embodiments, the total number of advertisement sources selected needs to meet certain conditions. In this example, the number of advertisement sources selected needs to be greater than the number of advertisement sources used to match the keyword set, since multiple advertisement sources can correspond to one keyword set. In some embodiments, the system selects twice as many ad sources as needed for keyword matching. However, in general, given the search speed and performance of an online commerce platform server, one keyword typically corresponds to one ad source or group of ad sources. Therefore, when searching for the advertisement source based on the keyword later, it is not necessary to traverse all the advertisement sources. Each keyword can search a different advertisement source, thereby increasing the search speed of the online commerce platform server. For example, the advertisement source “sales of various mobile phones” is associated with the keyword set “mobile phone”, and the advertisement source “sales of mobile phones” is also associated with the keyword set “mobile phone”. Two advertising sources and their matching relationships can coexist so that they can correspond to the same set of keywords.

After selecting the advertising source provided by the seller for keyword matching, a planning model is established at step 304 with the goal of maximizing the seller's profit. In particular, an n-by-m matrix X is established. The rows of the matrix correspond to the keyword set and the columns correspond to the advertisement source. Therefore, the matrix component X _{i, j} indicates whether the i-th keyword matches the j-th advertisement source, that is, whether there is a correspondence between the i-th keyword and the j-th advertisement source. Where i = 1,..., N, and j = 1,.

  In step 305, the matrix is converted to a vector, and a 0-1 type integer programming model is established based on the vector after the conversion.

ここで、Ｐ_ijはｉ番目のキーワードがｊ番目の広告ソースに適合する場合の利益を表し、Ｘ_ijはｉ番目のキーワードがｊ番目の広告ソースに適合するか否かについて、１で適合を表し、０で非適合を表し、Ｎ_jはｊ番目の広告ソースが持ちうるキーワードの最大数を表し、ｎはキーワードの総数を表し、ｍは選択された広告ソースの総数を表す。 To convert from a matrix to a vector, the second column of the matrix is joined after the first column, the third column is joined after the second column, and so on, and so on. The last column is joined after the penultimate column. Thus, the matrix is converted into a vector having n * m components. The planning model establishes the following objective function based on the vectors and constraints:

Here, P _ij represents the profit when the i-th keyword matches the j-th advertisement source, and X _ij determines whether the i-th keyword matches the j-th advertisement source by 1 N _j represents the maximum number of keywords that the j th advertisement source can have, n represents the total number of keywords, and m represents the total number of selected advertisement sources.

  The objective function is solved using search-based techniques to find a solution for the match between the selected ad source and the keyword set that provides maximum seller profit under constraints.

  The match between the i th keyword and the j th advertisement source is represented by a set of values. In some embodiments, advertising source profit, similarity, advertising source width, and advertising source depth are used to represent the fit. In step 306, the advertising source profit, similarity, advertising source width, and advertising source depth corresponding to each match between each keyword and the advertising source are obtained.

  In this example, the profit, similarity, advertisement source width, and advertisement source depth corresponding to the match between the i th keyword and the j th advertisement source of the matrix X are obtained. The width of the advertisement source represents the maximum range in which one keyword set fits the advertisement source. The depth of the advertisement source represents the maximum range in which one advertisement source fits the keyword set.

  In practice, profit, similarity, advertisement source width, and advertisement source depth are numerical values found in embodiments of the present application. The similarity represents the similarity between the advertisement source and the keyword, and can be calculated using a well-known technique.

In some embodiments, the width of the advertisement source and the depth of the advertisement source are calculated as follows:
Ad source width = (number of keywords restricted to one ad source)-(number of keywords currently matching the ad source),
Ad source depth = Number of keywords restricted to one ad source.

  In step 307, among the plurality of matching combinations that have the greatest profit for the advertising source, the seller's profit achieved by these matching combinations is each calculated.

Initially, all matching combinations in the vector are sorted according to the advertising source interest. The top matching combination with the greatest benefit for the advertising source is determined. The seller's profit from each top-level matching combination is calculated. The seller's profit is calculated as follows:
Seller's profit = (maximum profit from a particular ad source and keyword set match)-(seller loss caused by match).
Here, the seller's loss caused by a particular match is defined as the maximum profit from matching another ad source to the same keyword set.

Note that the seller's profit can be calculated using a valuation function, ie, the seller's profit from a particular fit is obtained by constructing the valuation function. In one example, the evaluation function is defined as follows:
Seller's profit from a specific match = (maximum profit from match between ad source and keyword)-(seller loss caused by match).

  At step 308, it is determined whether there are multiple matching combinations with the largest seller profit, and if so, the multiple matching combinations with the largest seller profit are similarities, ad source widths, and / or Sorted according to the depth of the advertising source, the top matching combination is selected from the final sorting result, and if it does not exist, one matching combination with the largest seller profit is directly obtained.

  In this step, if there is only one matching combination with the largest seller profit, one matching combination with the largest seller profit is obtained and the next step is executed. When there are a plurality of matching combinations having the maximum seller profit, the plurality of matching combinations having the maximum seller profit are sorted according to the similarity, and the matching combination having the maximum similarity is obtained according to the result of the sorting. Furthermore, if there is only one matching combination having the maximum similarity, one matching combination having the maximum seller profit is obtained, and if there are a plurality of matching combinations having the maximum similarity, the maximum similarity is obtained. Are matched according to the width of the advertisement source, and the top matching combination is obtained as a result of this process. In this step, the best matching combination relationships are obtained in the following order: profit, similarity, ad source width, and / or ad source depth.

  In step 309, it is determined whether or not the number of keywords that match the advertisement source in the currently acquired matching combination has reached the upper limit. If the upper limit has been reached, the matching of the matching combination is deleted, and if the upper limit has not been reached, the matching combination is stored.

  In this example, there is a predetermined constraint on the number of keywords that each advertising source is allowed to match, so in this step, the match obtained in the previous step is verified against this constraint. In other words, it is checked whether the number of keywords that match the advertising source in the match reaches a threshold; if the threshold is reached, the matching matching is deleted and the next step is performed; If not, the correspondence between the ad source and keywords indicated by the match is stored and it is required that the match cannot be searched again in the next traversal, so the match correspondence is also preferably deleted, and The following steps are performed.

  In step 310, matching combinations that satisfy the constraints continue to be obtained from other matches between the keyword set and the advertising source until all of the selected advertising sources match the upper limit number of keywords.

  After obtaining a match between an ad source and a keyword set, other matches between another keyword and another ad source are found by traversing the matrix. The approach of obtaining a match through traversal may be the same as steps 306-309 until all of the ad sources selected in step 302 match the upper limit number of keywords.

  In step 311, the acquired matching combinations are sorted according to the total profit of each advertisement source, and after the sorting, matching combinations that satisfy the constraints are considered as solutions.

  The total profit of the selected advertising source is then calculated. For example, if the advertising source matches two keywords, the total profit is equal to the sum of the profits when the advertising source matches each of the two keywords. The resulting fit combinations are then sorted according to gross profit, and a certain number of fit combinations are selected as solutions from the sort results. The number of matching combinations is equal to the number of advertising sources used for matching.

  Steps 201 and 202 of the process 200 are provided after the step 311 to retrieve the advertising source provided by the seller based on the match generated in the process 200 based on the keyword set entered by the potential buyer. May be. By using this search method, the displayed advertising source can be closer to the potential buyer's purchase intention while better covering the product areas offered by the seller. This eliminates the need for potential buyers to repeatedly enter keywords for search in commerce on the online commerce platform. This can reduce the number of searches of the online commerce platform server and improve the speed and performance of the search.

  For purposes of explanation, a detailed example is shown in connection with process 300. Assume that the advertisement source provided by a given seller has already been classified and the distribution of sources has been obtained according to processing steps 301 and 302.

  In step 303, an advertisement source provided by the seller that satisfies a predetermined condition is obtained in association with the category.

  In this example, the predetermined condition requires that at least three matching advertisement sources each have at least two keywords. In this example, there are 100 advertising sources provided by Seller A, and the selected advertising sources are labeled as 1, 2, 3, 4, 5, and 6.

  Next, a corresponding keyword set is generated from the acquired advertisement source. In this example, the keyword set generated by the advertisement source 1 is a, b and c, the keyword set generated by the advertisement source 2 is c, d and e, and the keyword set generated by the advertisement source 3 is d and e, the keyword set generated by the advertisement source 4 is a and d, the keyword set generated by the advertisement source 5 is a, d and e, and the keyword set generated by the advertisement source 6 is a, b and d.

In step 304, a matrix X is established. In this example, the rows of the matrix correspond to keywords and the columns of the matrix correspond to advertisement sources. The matrix component X _{i, j} indicates whether or not the i th keyword matches the j th advertisement source.

In step 305, the matrix is converted to a vector and the following planning model is established according to the vector.

  A planning model is established according to the transformed vector. In some embodiments, a substantially optimal result of the planning model objective function is calculated using a search-based solution with reference to the specific situation of the fit between each keyword and each advertising source. . In the following, the calculation process of the search-based solution will be described.

  In step 306, the advertising source profit, similarity, advertising source width, and advertising source depth corresponding to the match between each keyword and each advertising source are respectively obtained.

In this example, Table 1 shows the benefit of matching, similarity, advertisement source width, and advertisement source depth between each of the six advertisement sources and each of the five keywords. Here, the four consecutive numbers in each input box represent profit, similarity, advertisement source width, and advertisement source depth, respectively.

  In step 307, among a plurality of matching combinations having the largest advertising source profit, a plurality of seller profits for the sellers with the plurality of matching combinations are calculated.

  Matching combinations are sorted according to the advertising source interests. The matching combination with the largest advertising source benefit “5” includes: Matching of keyword a and advertising source 1: 5, 3, 2, 1, matching of keyword a and advertising source 5: 5, 2, 1, 4, matching of keyword a and advertising source 6: 5, 2, 1, 3, Matching between keyword c and advertisement source 1: 5, 3, 2, 4, Matching between keyword d and advertisement source 2: Matching between keyword 5, 3, 5, 3, keyword e and advertisement source 2: 5, 3, 3, 1.

The seller profit for the seller resulting from the matching combination for each keyword is calculated. Seller profit is calculated by subtracting the seller's loss caused by the match from the maximum profit of the match between the advertising source and the keyword. Thus, in this process, the seller profit for the seller with each fit is calculated as follows:
Matching between keyword a and advertising source 1: 5-5 = 0
Matching between keyword a and advertisement source 5: 5-5 = 0
Matching between keyword a and advertisement source 6: 5-5 = 0
Matching between keyword c and advertising source 1: 5-4 = 1
Matching between keyword d and advertisement source 2: 5-4 = 1
Matching between keyword e and advertising source 2: 5-4 = 1

  At step 308, the multiple matching combinations with the largest seller profit are re-sorted in order according to similarity, ad source width, and / or ad source depth, and the top result from the final sort result. A matching combination is selected.

  The matching combination with the largest seller profit is selected in order of seller profit in the previous process. The three matching relationships that provide maximum seller profit are c: 1, d: 2, and e: 2. Then further sorting according to the other three values yields: Matching of keyword d and advertising source 2: Matching of keyword e and advertising source 2: Matching of keyword e and advertising source 2: Matching of keyword c and advertising source 1: 1, 3, 2,4.

  In step 309, it is determined whether the number of keywords that the advertising source in the currently acquired match has reached the upper limit. When the upper limit is reached, the matching of the matching combination is deleted, and when the upper limit is not reached, the matching combination is stored.

  Since multiple matches have the same seller profit and similarity, the first match is chosen as the match between ad source 2 and keyword set d, but the 2-d match is compared to the remaining matches Has the maximum width value. It is further determined whether or not the number of keywords that the advertising source 2 matches has reached a limit (2 in this case). If the limit has not been reached, the match is stored, and if the limit has been reached, the match between ad source 2 and keyword set d is deleted so that the same match will not be found again later when the process is repeated. (Ie, all corresponding values for keyword d and ad source 2 are set to 0).

  In step 310, the resulting matching combinations are sorted according to the total profit of each advertising source, and after sorting, the matching combinations that satisfy the constraints are considered solutions.

  In step 311, all matching combinations obtained for the six ad sources are sorted according to the total profit of the six ad sources, ie the sum of the profits for each of the two keywords corresponding to the six ad sources. . The top three sorted advertising sources are then considered solutions. The solution may also be returned to the seller as a match between the advertising source and the keyword.

  In practice, a greedy algorithm may be used to make a selection within the matching combination that has the greatest seller benefit given the performance of the algorithm. Depending on the actual situation, different algorithms and implementation processes may be used. Because the search for the advertising source of the present application is based on the specific fits generated by the present application, the searched advertising source will allow potential buyers to search for a more accurate advertising source with fewer queries. Therefore, the number of searches by the online commerce platform can be reduced, and the search performance of the online commerce platform can be improved.

  For simplicity of explanation, the above-described method embodiments are described as a combination of a series of operations. However, it should be understood by one of ordinary skill in the art that the present application is not limited to the order of operations described, as some of the steps may be performed in another order or simultaneously in accordance with the present application. Further, it is to be understood that the embodiments presented herein are preferred embodiments and that the operations and modules associated therewith are not essential to the practice of the present application.

  FIG. 4 is a block diagram illustrating one embodiment of a device for generating a match between an advertisement source and a keyword. In this example, the device includes a selection module 401 adapted to select an advertisement source provided by the seller that meets a predetermined condition in relation to the category, and a corresponding keyword from the selected advertisement source provided by the seller. Establishing a planning model that includes a keyword generation module 402 adapted to generate and an objective function subject to predetermined constraints and represents a match between the selected advertising source provided by the seller and the generated keyword A planning model establishment module 403 adapted to obtain a fit, and a fit acquisition module 404 adapted to solve an objective function using a search-based solution to obtain a fit between the advertisement source and the matched keyword .

  FIG. 5 is a block diagram illustrating one embodiment of a search system in an online commerce platform. The system includes an adaptation generation unit 503. The adaptation generation unit 503 selects an advertisement source provided by the seller that satisfies a predetermined condition in relation to the category, and responds from the selected advertisement source provided by the seller. To generate a keyword that includes the objective function subject to the given constraints, establishes a planning model that represents a match between the selected ad source provided by the seller and the generated keyword, Is adapted to solve the objective function using a search-based solution to obtain a match between the generated keywords. The system further includes a search unit 501 adapted to search for a corresponding advertisement source according to a match based on keywords received and received by the potential buyer, and the retrieved corresponding advertisement source to the potential buyer. A display unit 502 adapted to display on a website page. The display unit may display audio, video, or any other suitable information that can be perceived by a potential buyer.

  The device provided in this embodiment may be integrated into the search engine server or connected to the search engine server as a separate device. When the method of the present application is implemented in software, it may be implemented as a function newly added to the search engine server, or may be implemented as a separate component. This application is not limited to the implementations of the described methods and devices.

FIG. 6 is a block diagram illustrating another embodiment of a system for generating a match between an advertisement source and a keyword. In this example, the device is provided by the seller under each category, with a classification module 601 adapted to classify all advertising sources provided by the seller in the current online commerce platform system according to predetermined classification criteria. A distribution acquisition sub-module 602 adapted to obtain a distribution of all advertisement sources, a selection sub-module 603 adapted to select a predetermined number of advertisement sources under each category according to the distribution, wherein A matrix establishment sub-module 604 adapted to establish a matrix X, wherein the predetermined number is greater than the number of adapted advertising sources, where the matrix component X _{i, j} is the j-th keyword Converts a matrix into a vector by column, indicating whether it matches the ad source of A transformation sub-module 605 adapted to establish a planning model with the following objective function according to:

  The system further includes parameter acquisition subs adapted to obtain the profit, similarity, ad source width, and ad source depth corresponding to the match between each keyword and each ad source, respectively. A module 606 is provided.

  The system further includes a calculation sub-module 607 adapted to calculate a plurality of seller profits from a plurality of matching combinations for the seller among a plurality of matching combinations having the greatest advertising source profit, and a maximum seller profit. Determining whether there is a plurality of matching combinations having, and if so, determining the plurality of matching combinations having the greatest seller profit according to the similarity, the width of the advertising source, and / or the depth of the advertising source Sort to select the highest matching combination in the final sorting result, and if not, the first one adapted to directly obtain one matching combination with the highest seller profit A conformance acquisition sub-module 608.

  The system further determines whether or not the number of keywords that match the advertising source in the currently acquired matching combination has reached the upper limit, and if the upper limit has been reached, matches the matching combination. If the upper limit is not reached, the determination sub-module 609 adapted to store the matching combination is provided.

  The system also continuously obtains matching combinations that satisfy the constraints from other matches between each of the keywords and each of the advertising sources until all the selected advertising sources match the maximum number of keywords. A second adaptation acquisition sub-module 610 adapted to do so.

  The system further comprises an ordering sub-module 611 adapted to sort the acquired matching combinations according to the total profit of each of the advertising sources and to solve the matching combinations that satisfy the constraints after sorting.

  The modules and / or units described above are hardware such as programmable logic devices and / or application specific integrated circuits designed to perform specific functions as software components running on one or more general purpose processors. Or as a combination thereof. In some embodiments, the module is a non-volatile storage medium (optical disk) such as a plurality of instructions for causing a computing device (personal computer, server, network device, etc.) to perform the methods described in the embodiments of the present invention. , A flash storage device, a portable hard disk, etc.) may be embodied in the form of a software product. Modules and / or units may be implemented on a single device or distributed across multiple devices. The functions of the modules and units may be integrated with each other or further divided into a plurality of submodules or subunits.

  Finally, related terms such as first and second are only used to distinguish one thing or action from another, or the actual relationship between these things or actions or Note that there is no claim or suggestion of order. Further, since a “comprising”, “including”, or variant thereof is intended for non-exclusive inclusion, a process, method, article, or apparatus comprising a series of elements includes these elements. It includes not only elements that are explicitly listed, but also other elements including elements that are originally provided in a process, method, article, or apparatus. Unless further limitations are given, an element defined by the expression “comprising (singular element)” means that there is an additional identical element in the process, method, article, or apparatus that includes the element. Is not to be excluded.

として確立すること、を含み、
Ｐ _ij は、前記ｉ番目のキーワードおよび前記ｊ番目の広告ソースの間の適合による売り手利益を表し、Ｘ _ij は、前記ｉ番目のキーワードが前記ｊ番目の広告ソースと適合するか否かを表し、Ｎ _j は、前記ｊ番目の広告ソースが持つことを許されるキーワードの最大数を表し、ｎは、前記キーワードの総数を表し、ｍは、前記選択済み広告ソースの総数を表す、方法。
適用例１１：適用例１０に記載の方法であって、前記目的関数は、検索ベースの解法を用いて解かれる、方法。
適用例１２：適用例１０に記載の方法であって、前記目的関数を解くことは、前記１組の所定の制約条件を満たす適合組み合わせを取得することであって、前記キーワードの各々と前記選択済み広告ソースの各々との間の適合に対応する広告ソースの利益、類似度、広告ソースの幅、および、広告ソースの深さをそれぞれ取得し、 最大の広告ソースの利益を有する複数の適合組み合わせの中で、前記売り手に対する前記複数の適合組み合わせから複数の売り手利益を計算し、最大の売り手利益を有する１または複数の適合組み合わせを決定し、現行の適合組み合わせにおける前記広告ソースが適合するキーワードの数が前記最大数に達しているか否かを判定し、現行の適合組み合わせにおける前記広告ソースが適合するキーワードの数が前記最大数に達している場合に前記適合組み合わせの対応付けを削除し、達していない場合に前記適合組み合わせを格納すること、を含む、前記１組の所定の制約条件を満たす適合組み合わせを取得することと、前記選択済み広告ソースすべてが前記最大数のキーワードに適合するまで、前記キーワードの各々と前記広告ソースの各々との間の他の適合から前記１組の所定の制約条件を満たす適合組み合わせを繰り返し取得し、前記広告ソースの各々の総利益に従って、前記取得した適合組み合わせをソートし、前記取得した適合組み合わせから前記１組の所定の制約条件を満たす適合組み合わせを前記目的関数の解として選択すること、を備える、方法。
適用例１３：オンライン広告システムであって、１または複数のプロセッサであって、売り手によって提供された複数の広告ソースの中から、所定の条件を満たす複数の選択済み広告ソースを選択し、前記複数の選択済み広告ソースに対応する複数のキーワードセットを生成し、前記複数の選択済み広告ソースおよび前記複数のキーワードセットの適合を表す計画モデルを１組の所定の制約条件に従って確立し、前記複数の広告ソースの少なくともいくつかと、前記複数のキーワードセットとの間の実質的に最適な適合を、前記計画モデルの解を求めることによって決定することを実行するように構成されている１または複数のプロセッサと、前記１または複数のプロセッサに接続され、前記プロセッサに命令を提供するよう構成されている１または複数のメモリと、を備える、システム。
適用例１４：適用例１３に記載されたシステムであって、前記プロセッサは、さらに、
潜在的な買い手によって入力されたキーワードセットを受信し、前記受信されたキーワードセットに基づいた前記実質的に最適な適合に従って対応する広告ソースを検索するように構成されている、システム。
適用例１５：適用例１３に記載のシステムであって、前記プロセッサは、さらに、前記検索された対応する広告ソースを前記潜在的な買い手に対してウェブページで表示するように構成されている、システム。
適用例１６：適用例１３に記載のシステムであって、前記プロセッサは、さらに、所定の分類基準に従って、前記売り手によって提供された前記複数の広告ソースを分類するように構成されている、システム。
適用例１７：適用例１６に記載のシステムであって、前記所定の条件を満たす前記複数の選択済み広告ソースの選択は、分類された１組の広告ソースと関連して実行される、システム。
適用例１８：適用例１７に記載のシステムであって、カテゴリと関連して、前記所定の条件を満たす前記売り手によって提供された前記広告ソースを選択することは、カテゴリに従って、前記売り手によって提供された前記複数の広告ソースの分布を取得し、前記分布に従ってそれぞれのカテゴリから所定の数の広告ソースを選択すること、を含み、前記所定の数は、キーワードセットと適合される広告ソースの数に対する売り手指定の制約条件よりも大きい、システム。
適用例１９：適用例１３に記載のシステムであって、前記制約条件は、キーワードセットと適合される前記広告ソースの数を含む、システム。
適用例２０：適用例１３に記載のシステムであって、前記制約条件は、前記広告ソースの各々に対するキーワードセットの最大数を含む、システム。
適用例２１：適用例１３に記載のシステムであって、前記１組の所定の制約条件に従って前記目的関数を確立することは、行列Ｘを確立し、前記行列の成分Ｘ _i,j は、ｉ番目のキーワードがｊ番目の広告ソースに適合するか否かを表し、前記目的関数を

として確立すること、を含み、
Ｐ _ij は、前記ｉ番目のキーワードおよび前記ｊ番目の広告ソースの間の適合による売り手利益を表し、Ｘ _ij は、前記ｉ番目のキーワードが前記ｊ番目の広告ソースと適合するか否かを表し、Ｎ _j は、前記ｊ番目の広告ソースが持つことを許されるキーワードの最大数を表し、ｎは、前記キーワードの総数を表し、ｍは、前記選択済み広告ソースの総数を表す、システム。
適用例２２：適用例２１に記載のシステムであって、前記目的関数は、検索ベースの解法を用いて解かれる、システム。
適用例２３：適用例２１に記載のシステムであって、前記目的関数を解くことは、前記１組の所定の制約条件を満たす適合組み合わせを取得することであって、前記キーワードの各々と前記選択済み広告ソースの各々との間の適合に対応する広告ソースの利益、類似度、広告ソースの幅、および、広告ソースの深さをそれぞれ取得し、最大の広告ソースの利益を有する複数の適合組み合わせの中で、前記売り手に対する前記複数の適合組み合わせから複数の売り手利益を計算し、最大の売り手利益を有する１または複数の適合組み合わせを決定し、現行の適合組み合わせにおける前記広告ソースが適合するキーワードの数が前記最大数に達しているか否かを判定し、現行の適合組み合わせにおける前記広告ソースが適合するキーワードの数が前記最大数に達している場合に、前記適合組み合わせの対応付けを削除し、達していない場合に、前記適合組み合わせを格納すること、を含む、前記１組の所定の制約条件を満たす適合組み合わせを取得することと、前記選択済み広告ソースすべてが前記最大数のキーワードに適合するまで、前記キーワードの各々と前記広告ソースの各々との間の他の適合から前記１組の所定の制約条件を満たす適合組み合わせを繰り返し取得し、前記広告ソースの各々の総利益に従って、前記取得した適合組み合わせをソートし、前記取得した適合組み合わせから前記１組の所定の制約条件を満たす適合組み合わせを前記目的関数の解として選択すること、を備える、システム。
Although the embodiments described above have been described in some detail for ease of understanding, the invention is not limited to the details provided. There are many alternative ways of implementing the invention. The disclosed embodiments are illustrative and not intended to be limiting.
Application Example 1: Online advertising method, selecting a plurality of selected advertisement sources satisfying a predetermined condition from a plurality of advertisement sources provided by a seller, and corresponding to the plurality of selected advertisement sources. Generating a plurality of keyword sets, establishing a planning model representing the plurality of selected ad sources and the match of the plurality of keyword sets according to a set of predetermined constraints, and at least some of the plurality of ad sources; Determining a substantially optimal fit between a plurality of keyword sets by determining a solution of the planning model.
Application Example 2: The method according to Application Example 1, wherein the planning model includes an objective function that is conditional on the set of predetermined constraints, and obtaining a solution of the planning model is a solution of the objective function. Including performing a search to seek.
Application Example 3: The method described in Application Example 1 further receives a keyword set entered by a potential buyer and searches for a corresponding advertisement source according to the substantially optimal fit based on the received keyword set. A method comprising:
Application Example 4: The method of Application Example 1 further comprises displaying the retrieved corresponding advertisement source on a web page to the potential buyer.
Application Example 5: The method of Application Example 1 further comprises classifying the plurality of advertising sources provided by the seller according to predetermined classification criteria.
Application Example 6: The method according to Application Example 5, wherein the selection of the plurality of selected advertisement sources that satisfy the predetermined condition is performed in association with a set of classified advertisement sources.
Application Example 7: The method according to Application Example 6, wherein selecting the advertisement source provided by the seller that satisfies the predetermined condition in relation to a category is provided by the seller according to a category. Obtaining a distribution of the plurality of advertisement sources and selecting a predetermined number of advertisement sources from each category according to the distribution, wherein the predetermined number is relative to the number of advertisement sources matched with the keyword set. Greater than the seller-specified constraints.
Application Example 8: The method of Application Example 1, wherein the set of predetermined constraints includes the number of advertisement sources that are matched with a keyword set.
Application Example 9: The method of Application Example 1, wherein the set of predetermined constraints includes a maximum number of keyword sets for each of the advertisement sources.
Application Example 10: The method according to Application Example 1, wherein establishing the planning model according to the set of predetermined constraints constructs a matrix X, and the components X _{i, j} of the matrix are i An objective function that represents whether the th keyword matches the j th advertisement source and is conditional on the set of predetermined constraints.

Establishing as
P _ij represents the seller profit due to the match between the i th keyword and the j th advertisement source, and X _ij represents whether the i th keyword matches the j th advertisement source. , N _j represents the maximum number of keywords that the j th advertisement source is allowed to have, n represents the total number of the keywords, and m represents the total number of the selected advertisement sources.
Application Example 11: The method according to Application Example 10, wherein the objective function is solved using a search-based solution.
Application Example 12: The method according to Application Example 10, wherein the solving of the objective function is to obtain a matching combination satisfying the set of predetermined constraints, and each of the keywords and the selection Multiple matching combinations with the maximum advertising source benefit, each obtaining the advertising source profit, similarity, ad source width, and ad source depth corresponding to the matching between each Calculating a plurality of seller profits from the plurality of matching combinations for the seller, determining one or more matching combinations having the largest seller profit, and determining the keyword source to which the advertising source in the current matching combination matches. Determine whether the number has reached the maximum number, and the number of keywords that the ad source matches in the current matching combination is Obtaining a matching combination that satisfies the set of predetermined constraints, including deleting the matching of the matching combinations when the maximum number is reached, and storing the matching combinations when not reaching the maximum number A matching combination that satisfies the set of predetermined constraints from other matches between each of the keywords and each of the advertising sources until all of the selected advertising sources match the maximum number of keywords. Repeatedly acquiring, sorting the acquired matching combinations according to the total profit of each of the advertising sources, and selecting from the acquired matching combinations a matching combination that satisfies the one set of predetermined constraints as a solution of the objective function Comprising a method.
Application Example 13: Online advertising system, which is one or a plurality of processors, and selects a plurality of selected advertisement sources satisfying a predetermined condition from a plurality of advertisement sources provided by a seller. Generating a plurality of keyword sets corresponding to a plurality of selected advertising sources, establishing a planning model representing the plurality of selected advertising sources and the matching of the plurality of keyword sets according to a set of predetermined constraints, One or more processors configured to perform a determination of a substantially optimal match between at least some of the advertising sources and the plurality of keyword sets by determining a solution of the planning model Connected to the one or more processors and configured to provide instructions to the processors Or comprising a plurality of memory, the system.
Application Example 14: The system described in Application Example 13, wherein the processor further includes:
A system configured to receive a keyword set entered by a potential buyer and search for a corresponding advertisement source according to the substantially optimal fit based on the received keyword set.
Application Example 15: The system of Application Example 13, wherein the processor is further configured to display the retrieved corresponding advertisement source on a web page to the potential buyer. system.
Application 16: The system of Application 13, wherein the processor is further configured to classify the plurality of advertising sources provided by the seller according to predetermined classification criteria.
Application Example 17: The system according to Application Example 16, wherein the selection of the plurality of selected advertisement sources satisfying the predetermined condition is performed in association with a set of classified advertisement sources.
Application Example 18: The system according to Application Example 17, wherein selecting the advertisement source provided by the seller that satisfies the predetermined condition in relation to a category is provided by the seller according to a category. Obtaining a distribution of the plurality of advertisement sources and selecting a predetermined number of advertisement sources from each category according to the distribution, wherein the predetermined number is relative to the number of advertisement sources matched with the keyword set. The system is larger than the restrictions specified by the seller.
Application example 19: The system according to application example 13, wherein the constraint includes the number of the advertisement sources matched with a keyword set.
Application example 20: The system according to application example 13, wherein the constraint includes a maximum number of keyword sets for each of the advertisement sources.
Application Example 21: The system according to Application Example 13, wherein establishing the objective function according to the set of predetermined constraints establishes a matrix X, and the components X _{i, j} of the matrix are i Indicates whether the th keyword matches the j th ad source, and the objective function is

Establishing as
P _ij represents the seller profit due to the match between the i th keyword and the j th advertisement source, and X _ij represents whether the i th keyword matches the j th advertisement source. , N _j represents the maximum number of keywords that the j th advertisement source is allowed to have, n represents the total number of the keywords, and m represents the total number of the selected advertisement sources.
Application example 22: The system according to application example 21, wherein the objective function is solved using a search-based solution.
Application Example 23: The system according to Application Example 21, wherein the solving of the objective function is to obtain a matching combination that satisfies the one set of predetermined constraint conditions, and each of the keywords and the selection Multiple matching combinations with the greatest advertising source benefit, each obtaining the advertising source profit, similarity, ad source width, and ad source depth corresponding to the matching between each of the completed ad sources Calculating a plurality of seller profits from the plurality of matching combinations for the seller, determining one or more matching combinations having the largest seller profit, and determining the keyword source to which the advertising source in the current matching combination matches. Determine whether the number has reached the maximum number, and the number of keywords that the ad source matches in the current matching combination is A matching combination that satisfies the one set of predetermined constraints, including: deleting the matching of the matching combination when the maximum number is reached, and storing the matching combination when the maximum number is not reached Acquiring and satisfying the set of predetermined constraints from other matches between each of the keywords and each of the advertisement sources until all of the selected ad sources match the maximum number of keywords It repeatedly obtains matching combinations, sorts the acquired matching combinations according to the total profit of each of the advertising sources, and finds a matching combination satisfying the one set of predetermined constraints from the acquired matching combinations. Selecting as a system.

Claims (20)

A method of online advertising executed by a computer ,
Select multiple selected ad sources that meet certain criteria from multiple ad sources provided by the seller,
Generating a plurality of keyword sets corresponding to the plurality of selected ad sources;
Establishing a planning model representing a match of the plurality of selected ad sources and the plurality of keyword sets according to a set of predetermined constraints;
Determining a substantially optimal fit between at least some of the plurality of advertising sources and the plurality of keyword sets by solving a solution of the planning model;
Equipped with a,
Establishing the planning model according to the set of predetermined constraints,
Construct a matrix X, and the component X _{i, j} of the matrix represents whether the i th keyword matches the j th advertising source,
An objective function subject to the set of predetermined constraints

As established,
Including
P _ij represents the seller profit due to the match between the i th keyword and the j th advertisement source, and X _ij represents whether the i th keyword matches the j th advertisement source. , N _j represents the maximum number of keywords that the j th advertisement source is allowed to have, n represents the total number of the keywords, and m represents the total number of the selected advertisement sources . The method of claim 1 further comprises:
Receive keyword sets entered by potential buyers,
Searching for a corresponding advertisement source according to the substantially optimal fit based on the received keyword set;
A method comprising: The method of claim 1 further comprises:
Displaying the retrieved corresponding advertisement source on a web page to the potential buyer. The method of claim 1 further comprises:
Classifying the plurality of advertising sources provided by the seller according to predetermined classification criteria. The method of claim 4 , wherein the selection of the plurality of selected advertisement sources that meet the predetermined condition is performed in association with a set of classified advertisement sources. 6. The method of claim 5 , wherein selecting the advertising source provided by the seller that meets the predetermined condition in relation to a category comprises:
Obtaining a distribution of the plurality of ad sources provided by the seller according to categories,
Selecting a predetermined number of advertising sources from each category according to the distribution;
Including
The method, wherein the predetermined number is greater than a seller specified constraint on the number of advertising sources matched with the keyword set.   The method of claim 1, wherein the set of predetermined constraints includes a number of the advertising sources that are matched with a keyword set.   The method of claim 1, wherein the set of predetermined constraints includes a maximum number of keyword sets for each of the advertising sources. The method of claim 1 , wherein the objective function is solved using a search-based solution. The method of claim 1 , wherein solving the objective function comprises:
Obtaining a conforming combination that satisfies the set of predetermined constraints,
Obtaining an ad source profit, similarity, ad source width, and ad source depth corresponding to a match between each of the keywords and each of the selected ad sources, respectively;
Calculating a plurality of seller profits from the plurality of matching combinations for the seller among the plurality of matching combinations having the greatest advertising source profit;
Determine one or more matching combinations with maximum seller profit,
Determine whether the number of keywords matched by the ad source in the current matching combination has reached the maximum number,
Deleting the matching of the matching combination when the number of keywords that the advertising source matches in the current matching combination reaches the maximum number, and storing the matching combination when the number does not reach,
Obtaining a matching combination that satisfies the set of predetermined constraints, including:
Until all of the selected advertising sources match the maximum number of keywords, repeatedly obtain matching combinations that satisfy the set of predetermined constraints from other matches between each of the keywords and each of the advertising sources. And
Sorting the acquired matching combinations according to the total profit of each of the advertising sources;
Selecting, as the solution of the objective function, a matching combination that satisfies the set of predetermined constraints from the acquired matching combination;
A method comprising: An online advertising system,
One or more processors,
Select multiple selected ad sources that meet certain criteria from multiple ad sources provided by the seller,
Generating a plurality of keyword sets corresponding to the plurality of selected ad sources;
Establishing a planning model representing a match of the plurality of selected ad sources and the plurality of keyword sets according to a set of predetermined constraints;
1 or configured to perform a determination of a substantially optimal match between at least some of the plurality of advertising sources and the plurality of keyword sets by determining a solution of the planning model Multiple processors,
One or more memories connected to the one or more processors and configured to provide instructions to the processors;
Equipped with a,
Establishing the planning model according to the set of predetermined constraints,
Establishing a matrix X, wherein the matrix component X _{i, j} represents whether the i th keyword matches the j th advertising source,
Objective function

As established,
Including
P _ij represents the seller profit due to the match between the i th keyword and the j th advertisement source, and X _ij represents whether the i th keyword matches the j th advertisement source. , N _j represents the maximum number of keywords that the j th advertisement source is allowed to have, n represents the total number of the keywords, and m represents the total number of the selected advertisement sources . The system of claim 11 , wherein the processor further comprises:
Receive keyword sets entered by potential buyers,
A system configured to search for a corresponding advertisement source according to the substantially optimal match based on the received keyword set. The system of claim 11 , wherein the processor is further configured to display the retrieved corresponding advertisement source on a web page to the potential buyer. 12. The system of claim 11 , wherein the processor is further configured to classify the plurality of advertising sources provided by the seller according to predetermined classification criteria. 15. The system of claim 14 , wherein the selection of the plurality of selected advertisement sources that meet the predetermined condition is performed in association with a set of classified advertisement sources. 16. The system of claim 15 , wherein selecting the advertisement source provided by the seller that meets the predetermined condition in relation to a category comprises:
Obtaining a distribution of the plurality of ad sources provided by the seller according to categories,
Selecting a predetermined number of advertising sources from each category according to the distribution;
Including
The system wherein the predetermined number is greater than a seller specified constraint on the number of advertising sources matched with the keyword set. 12. The system of claim 11 , wherein the constraint includes a number of the advertising sources that are matched with a keyword set. The system of claim 11 , wherein the constraint includes a maximum number of keyword sets for each of the advertising sources. 12. The system according to claim 11 , wherein the objective function is solved using a search-based solution. 12. The system of claim 11 , wherein solving the objective function is
Obtaining a conforming combination that satisfies the set of predetermined constraints,
Obtaining an ad source profit, similarity, ad source width, and ad source depth corresponding to a match between each of the keywords and each of the selected ad sources, respectively;
Calculating a plurality of seller profits from the plurality of matching combinations for the seller among the plurality of matching combinations having the greatest advertising source profit;
Determine one or more matching combinations with maximum seller profit,
Determine whether the number of keywords matched by the ad source in the current matching combination has reached the maximum number,
Deleting the matching of matching combinations when the number of keywords that the advertising source matches in the current matching combination has reached the maximum number, and storing the matching combination if not,
Obtaining a matching combination that satisfies the set of predetermined constraints, including:
Until all of the selected advertising sources match the maximum number of keywords, repeatedly obtain matching combinations that satisfy the set of predetermined constraints from other matches between each of the keywords and each of the advertising sources. And
Sorting the acquired matching combinations according to the total profit of each of the advertising sources;
Selecting, as the solution of the objective function, a matching combination that satisfies the set of predetermined constraints from the acquired matching combination;
A system comprising:

Images